From data to information - De Zoogdiervereniging

Page created by Danny Peterson
 
CONTINUE READING
From data to information - De Zoogdiervereniging
From data to information

A start document for setting up a functional
blueprint that makes data generated with
automatic bat detectors available for a broad
use

Limpens, H.J.G.A, V.J.A. Hommersen and M.J. Schillemans

   2019.29
   Report of the Dutch Mammal Society
   Commissioned by Netherlands Biodiversity Information
   Facility
From data to information

A start document for setting up a functional
blueprint that makes data generated with
automatic bat detectors available for a broad use
 Rapport nr.:                              2019.29

 Datum uitgave:                            Februari 2020

 Status                                    Definitief

 Auteur:                                   Limpens, H.J.G.A, V.J.A. Hommersen and M.J.
                                           Schillemans

 Kwaliteitscontrole:                       H.J.G.A. Limpens

 Productie:                                Steunstichting VZZ, in report regarded as
                                           Dutch Mammal Society
                                           Bezoekadres: Toernooiveld 1
                                                          6525 ED Nijmegen
                                           Postadres:     Postbus 6531
                                                          6503 GA Nijmegen
                                           Tel.: 024 7410500
                                           secretariaat@zoogdiervereniging.nl
                                           www.zoogdiervereniging.nl
 Gegevens opdrachtgever:                   Netherlands Biodiversity Information
                                           Facility
                                           Postbus 93102

 Contactpersoon opdrachtgever              Dr. Niels Raes. NLBIF, node manager

               Steunstichting VZZ is part of the Dutch Mammal Society

          Suggested citation:
          Limpens, H.J.G.A, V.J.A. Hommersen and M.J. Schillemans, 2020. From data to
          information. Rapport 2019.29. Bureau van de Zoogdiervereniging, Nijmegen.

  De Steunstichting VZZ, onderdeel van de Zoogdiervereniging, is niet aansprakelijk voor gevolgschade, alsmede voor schade welke
  voortvloeit uit toepassingen van de resultaten van werkzaamheden of andere gegevens verkregen van de Zoogdiervereniging;
  opdrachtgever vrijwaart de Stichting VZZ voor aanspraken van derden in verband met deze toepassing.
  Niets uit dit rapport mag worden verveelvoudigd en/of openbaar gemaakt worden d.m.v. druk, fotokopie, microfilm of op welke
  andere wijze dan ook, zonder voorafgaande schriftelijke toestemming van de opdrachtgever hierboven aangegeven en de
  Zoogdiervereniging, noch mag het zonder een dergelijke toestemming worden gebruikt voor enig ander werk dan waarvoor het is
  vervaardigd.

  © Zoogdiervereniging
Contents

1         Introduction ..................................................................................... 2
    1.1       Background .............................................................................................. 2
    1.2       Goals ....................................................................................................... 3

2         Methods............................................................................................ 5
    2.1       Literature research .................................................................................... 5
    2.2       Consultation by questionnaire ..................................................................... 5

3         Literature research ........................................................................... 6

4         Questionnaire ................................................................................... 9

5         Analysis of consultation by questionnaire ...................................... 10

6         Synthesis and advice ...................................................................... 12

7         References ..................................................................................... 16
    7.1       Websites and web based –bat- data collection programs ...............................17
    7.2       Other references ......................................................................................17

8         Bijlages .......................................................................................... 21
From data to information

1 Introduction

1.1 Background

Bat research has developed significantly in the past decades. New technologies
and the enhanced practicality of new generations of ultrasound or bat detectors
contributed to this development.

Bat detectors are all about making ultrasound audible and/or recordable.
Basically bat detectors have evolved from tuneable heterodyne detectors (narrow
band, HET), via frequency division (broad band, FD), time expansion (broad
band, TE) to real time ultrasound recording (RT). Especially the last technique
innovates bat research through the employment of bat detectors that use
automatic triggering and real time recording. Development of high speed
soundcards and/or high frequency digitizing of sound have made digital recording
in real time possible. Automated triggering has made the employment of stand-
alone bat detectors possible.

This automated RT equipment records ultrasonic sounds automatically in reaction
on a pre-set trigger (frequency + sound level at certain frequencies). The system
is not dependent on a human reacting to an ultrasound event with an active
recording, thus creating new opportunities for standardised methods for active
and long term data collecting regarding species presence and/or bat activity on a
global scale. Involvement of appropriately trained and coordinated volunteers in
surveys using broadband RT bat detectors creating increased species coverage
and objectivity of species identification, embody the future in developing
comprehensive bat monitoring programmes (Barlow et al., 2015). Automatic bat
detectors however, generate big amounts of data, especially when combined
with large scale utilization in e.g. citizen science projects.

Data storage
Globally there are several organisations/agencies/trusts that carry out citizen
science projects targeting bats. All these citizen science projects generate big
quantities of data. This big data availability creates new opportunities to
generate knowledge on biodiversity patterns. However, the large amount of data
also results in challenges in data handling. One main challenge is the data
storage. Where are such amounts of data stored? They can be stored on hard
drives, resulting in a risk if loss or damage of data. Storage online may reduce
such risks. It is, however, not clear where to store such large amounts of data.
And new challenges can be found in questions like: Which data to save of delete?
Who to allow access to the data? How to secure your data?

Data analysis
Another challenge is that the collected data files/sound files need to be analysed,
which is either being done more centralized by the organisations/agencies/trusts
that collate the data, or more decentralized by employing volunteers. When the
data is analysed by a central organisation this requires a large investment of
time by this organisation. On the other hand, this removes the potential

                    2
From data to information

influence of variation in volunteer experience and skills (Barlow et al., 2015;
Newson et al., 2015). Data analysed by volunteers create other difficulties, like
data not being analysed because of the great amount of time and effort
requested from the volunteers, the (relative) lack of experience of the (different)
volunteer(s). Training and coordination of volunteers, and the use of
standardised methods, have found to be crucial to the success of surveys with
volunteers (Newman et al., 2003).

Software
Besides that, the software dealing with processing sound files is developing
continuously and fast, ever creating new opportunities e.g. to speed up analysis.
Filtering raw data – recorded sound files - is becoming faster and more accurate.
There are, however, still (too) many sound files not allocated, or files that are
only identified to a genus or group level. It is therefore needed to share
experiences with partners that use these software.

Data fragmentation
Another challenge is the fragmentation of generated bat data. Because there is
no solid –global- data-infrastructure, data is only shared on a local scale. Most of
the time data is only available for project participants and/or is stored only on
the hardware of the observer. This creates a situation in which potentially
valuable data is not available or accessible for researchers and scientific analysis
and/or a wider audience. Also data from different projects is hardly shared.

The challenges regarding automatic generated data, e.g. generated by employing
RT recorders, are characteristic for a new generation of biodiversity data: data
that is automatically generated and produces big amounts of digital information.
Comparable cases are e.g. data generated by the GPS tracking of birds or
automatic generated identifications of plankton using plankton recorders. Al such
data need to be stored, be comparable with existing information systems and
also be usable in the future. The Global Biodiversity Information Facility (GBIF) is
globally the most important infrastructure of biodiversity data. An organisation
like GLBIF, and their national partners like NLBIF, respond to these new
developments, which may be expected to play an increasingly important role in
scientific biodiversity research. The approach of this report is therefore to get an
overview of the circumstances regarding (RT) bat data and to study which first
steps need to be taken to standardise the storage, analysis and interoperability
of these data to improve the use of it.

1.2 Goals

Our goals are:
1) To carry out an analysis of problems and opportunities,
2) with a team of people that have experience in processing big amounts of
data/analysing sounds files of bats,
3) in which bottlenecks in the data use, data processing and data presentation
will be identified and prioritized,
4) and the relevant questions and proposals of solutions of the bottlenecks will
be formulated.

                    3
From data to information

The goal of this project is to make a start document for setting up a functional
blueprint for a system for data handling that makes data generated with
automatic bat detectors available for a broader use. The data should be usable
for national and international researchers and volunteers via GLBIF and –where
applicable- the (Dutch) National Database Flora and Fauna (NDFF). The start
document aims at making progress in creating a usable dataflow that leads to
usable data. This would be the first step in setting up a data-infrastructure for
big amounts of bat data.

                    4
From data to information

2 Methods

The project is divided in four phases: literature research, consultation by
questionnaire, analysis of questionnaire and analysis and synthesis.

2.1 Literature research

In the literature research we studied worldwide initiatives and researches other
than the initiatives of the Dutch Mammal Society, regarding bats and big
amounts of data. By doing this we can get a first overview of the bottlenecks
people are facing regarding this topic. In addition, relevant persons are noted
that will be contacted for their input in the questionnaire.

2.2 Consultation by questionnaire

The analysis of challenges regarding bats and big amounts of data are further
addressed by consulting a maximum of ten researchers having experience with
these problems from different points of perspective. People are contacted by
questionnaire. The questionnaire mainly deals with (meta) data and data
retrieval and storage. The outcome of the response is analysed and described
(Chapter 4). This outcome was communicated to the people that gave input for
the questionnaire. The analysis of the questionnaire has been synthesised in
Chapter 5. This resulted in an advice and conclusions, that will be addressed in
Chapter 6.

                    5
From data to information

3 Literature research

Globally there are several organisations/agencies/trusts that carry out citizen
science projects regarding bats (Table 2). Apart from two already existing – non
acoustic - monitoring networks, the Dutch Mammal Society set up a monitoring
network based on acoustic observation of bats targeting population trends of four
bat species. Data is collected by volunteers, who each year carry out car (or
bike) transects. Species identification is also done by volunteers, who are trained
by means of yearly workshops.

Also in Ireland and France citizen science acoustic monitoring projects are carried
out, as Bat Conservation Ireland and VigieNature (France) organise car transects
for volunteers. In contrast to the Dutch Mammal society, Bat Conservation
Ireland analyses the data centralized without the help of volunteers (Roche et
al., 2005 and 2011). In France the recordings and identification results are
uploaded via a portal. Identification is done by volunteers using (custom)
software.

Concerning the UK, Bat Conservation Trust (BCT) has a National Bat Monitoring
Program with several ?? acoustic surveys, carried out by volunteers. Data
analysis is accomplished by the volunteers themselves, who are trained by
means of workshops and online training. BCT also has a partnership project with
The Institute of Zoology (IoZ): The Indicator Bats Program (iBats). It aims to
develop national bat monitoring programs across the globe (BCT, n.d.).
Recordings are provided to volunteers who then identify whether a bat is
present, and if so, what species it might be. The same recording is presented to
several volunteers. On the basis of this input (especially regarding the presence
of a bat in a recording) an algorithm is developed to extract bat calls, which then
in turn can be identified to species level (Mac Aodha et al., 2018).
       Other citizen science bat projects in the UK involve the Norfolk Bat Survey
and the Southern Scotland Bat Survey. These involves surveys in in Norfolk,
Southern Scotland and in the Norfolk and Suffolk Brecks, carried out by
volunteers. They borrow a bat detector from a Bat Monitoring Centre and collect
data in 1 km squares (Norfolk Bat Survey, n.d.; Southern Scotland Bat Survey,
n.d.). Initially data were stored on a SD-card and sent by post, but it is now also
possible to upload recordings directly. Identification of the species is done using
software with a fast feedback to the volunteers.

In America, the Wisconsin Bat Program organises land, water and driving
transects for volunteers. Data is saved onto the PDA and analysed in the office
(Wisconsin Bat Program, n.d.).
      Another bat monitoring program in North America is yet in progress. The
North American Bat Monitoring Program (NABat) states that there are currently
no national programs to monitor and track bat populations in North America. The
aim of the NABat is to provide the architecture for coordinated bat monitoring
and to provide managers and policy makers with the information they need on
bat population trends (USGS, 2016). Currently (NABat) is using acoustic data
gathered from various sources and sampling designs, named ‘ projects’ (Loeb at

                    6
From data to information

al., 2015). Projects may contain sampling form one to a various number of grid
cells. Species ID is done by the collector of the data using various types of
software. Results are uploaded using a project-database supplied by NABat,
collecting the species identification, parameters generated by the software used
to identify the species and – possible but not required- the actual recordings.

Data of the Norfolk Bat Survey and the Southern Scotland Bat Survey were
analysed by means of a first analysis with the Software SonoChiro
(http://www.biotope.fr/fr/innovation/sonochiro), a record filtering in SAS and
then manual checking using the software SonoBat (http://www.sonobat.com/) of
the files. SonoChiro provided robust results for the majority of the species and
recordings were filtered to remove low quality recordings and identifications with
low confidence (Newson et al 2015). Currently the data is being analysed using
custom software (Tadarida) (https://www.batsurvey.org/species-identification/).

Because of the fragmented data of population status and trends in America, the
USGS started in 1994 with the USGS Bat Population Data Project. In this project
bat data and publications were synthesized, data was tested for the utility for
estimating trends and evaluated for future monitoring programs. The projects
product was the bat population data base (USGS, 2017). Currently, USGS
scientists want to update , updating and extending the capabilities of this
database for better data management, accessibility and utility. Also Bat
Conservation International and Washington DC-based NatureServe have formed
partners to launch the initiative of a bat database, this is a global database.

The first international symposium on Bat Echolocation methods was organised in
2002 in Austin Texas. This meeting resulted in a handbook on the use of acoustic
methods, focussing on the techniques of observation and identification of bats in
the field with bat detectors (Brigham et al. 2004). In 2017 in Tucson Arizona the
second meeting was organized aiming at an update of the handbook (abstracts:
Fenton et al. 2017). The focus in the second meeting was on automated
recording and identification of bats. To date a revised version of the handbook is
not available.

                   7
From data to information

Table 2. A non-exhaustive Overview of organisations/agencies/trusts that carry
out acoustic citizen science surveys on bats and of bat researchers that work
with large datasets of bats.
                                                Names/
Initiatives       websites                      contacts         Description
iBats (Indicator http://www.bats.org.uk/pa Kit Stoner, Kate
 bats Program)     ges/ibatsprogram.html         Jones             Monitoring bats worldwide by means of transects
                                                              Monitoring bats throughout Norfolk by means of three
  Norfolk bat     http://www.batsurvey.org/                      survey locations in 1km-square. Detectors are
    survey                 norfolk/             Stuart Newson  borrowed and returned to Bat Monitoring Centres.
                                                                  Monitoring bats throughout Southern Scotland by
   Southern                                                        means of three survey locations in 1km-square.
  Scotland Bat    http://www.batsurvey.org/                       Detectors are borrowed, returned and analysed by
    survey                scotland/             Stuart Newson               southern Scotland bat survey.
                                                                 Monitoring bats throughout Suffolk and Breckland by
                  http://www.batsurvey.org/                        means of three survey locations in 1km-square.
 The Breckland            projects-in-                            Detectors are borrowed, returned and analysed by
  Bat Project       suffolk/breckland_bats/       James Parry                 the Breckland bat society.
  University of   http://biosciences.exeter.a
     Exeter,      c.uk/staff/index.php?web_i
   researcher           d=Paul_Lintott            Paul Lintott                         PhD on bats
     Bat
 Conservation     www.batconservationirelan                         Car based monitoring: 28 x 30km squares are
   Ireland                 d.org                 Tina Aughney                    surveyed by car.
                                                 Christian
                                              Kerbiriou, Jean
                  http://vigienature.mnhn.fr/ François Julien      Car based monitoring, counts in 2x2 km squares,
   Vigie Chiro          page/vigie-chiro        & Yves Bas                   stations in 2x2 km squares
                  http://wiatri.net/Inventory
 Wisconsin Bat    /Bats/volunteer/acoustic.cf
   Program                     m                                        Car, walking and canoe based transects
                                                                      Objectives: 1) provide the architecture for
North American                                                     coordinated bat monitoring to support inferences
Bat Monitoring                                                   about trends in bat populations and abundances, and
   Program     https://www.fort.usgs.gov/                            2) provide managers and policy makers with
   (NABat)         science-tasks/2457            Patty Stevens           information on bat population trends.
   USGS Bat
Population Data                                                   Creating one big database for bat data in the United
    Project       https://my.usgs.gov/bpd/        Bill Rainey                           States
                                                   Katherine
                                                   Boughey
      Bat                                        (Monitoring &
  conservation    http://www.bats.org.uk/pa         Science       Several monitoring projects, like iBats and National
     trust          ges/batmonitoring.html         Manager)                     Bat monitoring study.
      Bat         http://www.batcon.org/our     Dave Waldien,
 Conservation     -work/initiatives/launch-a-   Mylea Bayless,
 international        global-bat-database       Wynifred Frick                 Launch a Global Bat Database
  University of
     Naples                                      Danilo Russo                       Researcher of bats
 Swiss Federal
  Institute for
 Forest, Snow
and Landscape     http://www.wsl.ch/info/mit
   Research            arbeitende/obrist         Martin Obrist                      Researcher of bats
   Swiss Bat
  Bioacoustics     http://www.fledermaus-          Elias Bader
  Group SBBG               be.ch/                                                   Researcher of bats

                             8
From data to information

4 Questionnaire

The questionnaire is displayed in Appendix I and has been sent to 25 persons
dealing with automatic generated data of bats and large datasets worldwide. Of
these 25, eight persons (Table 3) gave their input for the questionnaire.

From the view point of data flow architecture, the questionnaire targeted
questions regarding what and how (meta)data is collected and retrieved and how
such data is finally stored. As a basis for a deeper level of interpretation of the
received feedback, also details where asked concerning recording, species
identification and monitoring/observation design.

Table 3. People that gave their input for the questionnaire.
 Initiatives/organisations/bat researchers      Name of person that gave input for
                                                questionnaire
 Norfolk bat survey / Southern Scotland Bat
 survey                                         Stuart Newson
 Vigie Chiro
                                                Yves Bas
 North American Bat Monitoring Program
 (NABat)                                        Brian Reichert
 Swiss Federal Institute for Forest, Snow and
 Landscape Research                             Martin Obrist
 NIOO-KNAW/ Dutch bat researcher
                                                Kamiel Spoelstra
 Dutch bat researcher
                                                Niels de Zwarte
 Dutch bat researcher
                                                Jasja Dekker
 Dutch bat researcher
                                                Sander Lagerveld

                        9
From data to information

5 Analysis of consultation by questionnaire

Out of the eight respondents three are dealing with nationwide and/or regional
data. Five respondents mainly deal with data concerning several (larger or
smaller) projects. Appendix II gives an overview. When we take the DMS
acoustic monitoring scheme into account there would be a fourth respondent
dealing with nationwide data and the number of respondents is then nine.

In the analysis we focus on four aspects relevant for dataflow architecture:
    1. Methods of transfer and storage of data
    2. Method of recording
    3. Identification of species
    4. Metadata

Method of transfer and storage of data
Data collected in the field is generally transferred to a central location by physical
means (e.g. hard drive, SD-card) or via file transfer. In four1 cases data is
transferred using a portal or interface for uploading. Mostly raw data is
transferred (e.g. recordings), in three cases a species ID is also transferred along
with the raw data.
       Data is mainly stored locally or via a SAN. When dealing with nationwide
monitoring schemes between 1 to 8 TB of storage per year is being used, and a
yearly increase (of unknown magnitude) is expected.

Method of recording
Automated bat detectors are always used. Brands of bat detectors and
microphones and their settings differ greatly. Across respondents different types
of triggers are used. Mostly full spectrum ‘time expanded’ or real time recordings
are being used.

Identification of species
When only recordings are transferred the identification process is obviously done
centrally. When species identification is done by the collectors of the data
(mainly volunteers) the software to do the identification is provided or freely
available. These are mainly software packages that run locally, although in
France the call extraction from the recordings is a local facility whose outcome is
then identified to species via an internet based tool. In the Norfolk scheme a
recent facility is offered where recordings are uploaded directly to a central
processing computer that generated a fast –partial- outcome of species, which is
fed back to the collector. Previously data was transferred by posting SD-cards.
       In all cases species identification is done using automated software. And in
almost all cases the result of the automatic identification software is validated by
hand vetting. Hand vetting is used on a sample of the data. The sample is either
chosen according to species group or to confidence level as indicated by the used
software. Sample size is not given.

1

                    10
From data to information

Identification mostly tries to target the species level, but this is not always
attainable. Several respondents group species together when identification to
species level is not possible. Most used are Myotis-group (genus), Nyctalus-
Serotinus-Vespertillio-group (guild) and the Plecotus-group (genus).

As recordings may also hold various non-bat sounds (artificial or other biota), a
process of scrubbing (extracting only the recordings with bat calls, or only bat
calls from recordings) is very important. A too strict process will also exclude
recordings with bat calls (false negatives), giving a false number of recordings
(or calls). A too loose process would give many ‘false positives’ which are falsely
regarded as bats (a species or a group).
        Among the respondents the process of scrubbing is either done using
settings of the recording device (triggers) of by filtering recordings afterwards.
There is no clear preference indicated by the respondents.

Metadata
Metadata which is retrieved alongside the actual recordings almost always
includes coordinates, date and time, used-IDs, data on the settings of the
recording device, data about the sampling point, data on call or recording quality
and specifics of used classifiers and filters.

                   11
From data to information

6 Synthesis and advice

From the literature research it is obvious that the use of automatic bat detectors
is growing and a lot of organisations face the related challenges in different
ways. We have to be aware that only a relatively small number of people
responded to the questionnaire and thus have to be careful when interpreting the
outcome.
       There is a marked difference in how respondents are dealing with data
between respondents running one or several, larger or smaller project(s) and
those who deal with nationwide of regional (monitoring)schemes).

Advancement of identification software is fast, as is the development of
automatic bat detectors. Both are decreasing in cost and growing in choice.
      This implies that the use of different makes of software and devices is also
growing. It is foreseeable that volunteers will increasingly use their own
automatic bat detector(s) to record bats while working in a monitoring scheme,
instead of using only one type which is provide by the project. This implies that
they can and will also collect data on an ad-hoc basis.
      At the same time, it seems clear that a recording system used nowadays
might be obsolete in a few years. This is a serious concern for strictly regulated
monitoring schemes, in which often – for the necessary methodical rigidity - only
one system of detection and recording is used.

Method of transferral and storage
There seems to be a trend towards uploading calls or recordings, instead of
physical transferral of data (e.g. hard disks, SD-cards), especially when dealing
with nationwide or regional monitoring schemes. However, one party mentioned
that the amount of data collected was growing too fast and they were thinking of
reversing to sending physical devices and/or doing local scrubbing (removing all
non-bats sound from the recordings) and then performing an upload.
       Also it is clear that when dealing with nationwide, or even continent-wide
data, a wish and need exists for a central database to store the data (including
metadata). Mostly data is stored on a SAN, which makes it easy to manage and
add storage capacity when needed. However, it is also foreseeable that when the
datasets become very large, searching such datasets will become more
strenuous. The storage facility will have to facilitate these searchers.

Method of recording
With strictly regulated monitoring schemes, mostly dealing with population
trends not based on occupancy methods, systems for detection and recording, as
well as the way they are used are also strictly regulated. But when dealing with
distribution data and ad-hoc data recordings systems and the way they were
employed will all differ. Using the full potential of distribution data will almost
always include dealing with different systems for detection and recording, modes
of employing and used trigger settings. This indicates the necessity of also
transferring metadata about devices, usage and settings.

                   12
From data to information

Identification
Species identification is always handled by software, rather than manually.
        Recently it has become clear that using software has important limitations
(among others Russo and Voigt, 2016). Although the accuracy of the software is
increasing and call extraction can be improved (for example see Mac Oadha et
al., 2018), errors will persist. These errors might be false positives (where a
recording is assigned a species identification while there is no bat pulse), false
negatives (a bat pulse is present in the recording, but no species identification is
assigned) or misidentification (a species identification is assigned but the
recording in is reality of a different species; in quantitative analysis these would
also add to the false negatives).
        Traditionally these errors are counteracted by hand vetting. This is indeed
indicated by all respondents. Hand vetting however involves much time,
depending on the sample taken to be hand vetted. The sample size is not given
by the respondents to the questionnaire. From literature it is clear that various
sample sizes are used. Sample sizes are based on a fixed percentage of all
recordings, a fixed percentage of recordings stratified towards species that are
difficult to identify using acoustics, a fixed percentage of recordings stratified
towards quality of recordings (e.g. S/N) or a mix of the before mentioned.

However, dealing with errors can also be based on the confidence level given by
the software used to identify the species in a recording, by setting an explicit
fault tolerance (Barré et al. 2019). Barré et al. (2019) show that the degree of
hand vetting needed thus can be significantly decreased. Furthermore they show
that when not taking errors into account analysis on ecological relationships
between species presence or abundance and environmental parameters or
landscape features, will lead to false conclusions.

This implies that when dealing with large datasets (of different sources) one
should always take the confidence level of the ID into account, and that it is
highly likely that this decreases the time needed for hand vetting. On turn, this
implies that metadata of the software used (e.g. settings and confidence levels)
should be available. It also implies that after recordings (raw data) and species
identifications are made available by observers, these metadata still needs to be
available for ID validation (hand vetting) and for dealing with errors and error
levels.

We can distinguish three modes to deal with the species ID process:
1) Decentralized: for example, NABat and the Dutch monitoring scheme.
   Collectors handle the species identification process themselves.
   Both recordings (raw data) and species ID (processed data) are transferred.
   This implies well trained collectors and transferral of metadata of the
   identification process (e.g. outcome of the automatic identification software
   including confidence levels and settings). It also implies that identification
   software is available for collectors.
   After recordings and identification data is retrieved, a validation process is
   needed to
   a) ensure identification is reliable and done with comparable assumptions

                   13
From data to information

   (e.g. confidence levels given by software, across different sources of data
   b) deal with expected errors of the software used.
   When dealing with many different sources of data (and thus also with many
   different recording devices and different software used for identification)
   metadata about the used devices and software becomes of paramount
   importance.
   When dealing with volunteers, a particular concern is that feedback should be
   presented regarding whether the identification supplied by the volunteer was
   indeed accurate. This ensures a learning effect. An important advantage of
   this decentralized fashion is that capacity is being generated among
   volunteers for dealing with acoustic data. This enables them to then also
   collect and report acoustic data on an ad-hoc fashion.

2) Centralized: for example, Ireland. Collectors send the recordings to a central
   location. There the identification process is handled. Feedback to collectors is
   done after processing the recordings. A major advantage is that the
   identification process can be tightly regulated and controlled. Making is easier
   to deal with errors. However, there is no or little learning effect for
   volunteers, which might also effect whether volunteers will stay involved in
   such schemes.

3) Mixed Mode: for example, France and UK. Collectors send the recordings to a
   central location. There the identification process is handled. In the case of
   France, collectors see output via an internet tool. In the UK, collectors receive
   a quick response concerning at least the common species and/or easily
   identified species and receive full information afterwards. As identification is
   done centrally is can be well controlled and regulated. There is no real
   learning effect, but volunteers receive feedback quickly, which is important to
   keep them involved.

A recent development is the collaboration between BTO, BCT, University College
London and Oxford University: together they try to develop and end-to-end
system of recording-automated identification-feedback
(https://www.bats.org.uk/our-work/national-bat-monitoring-programme/british-
bat-survey). An integral part of this system is also the automated feedback to
users.

Metadata
As discussed, metadata should not only deal with information about the
recording and sample site itself (location, data, time et cetera) but also with
information about the recording device and settings and if appropriate of the
identification process (software used, confidence levels of identification, settings
etc.)

                    14
From data to information

Implications for the Netherlands

   -   The evolution of automatic detectors results in an increasing number of
       volunteers using their private detector and/or recording device. These
       observers will use their personal equipment for regulated monitoring
       schemes as well as for their ‘personal projects’, leading to standardised as
       well as to ad-hoc data.
   -   There is an implicit wish for capacity building (learning).
   -   It is to be expected that a currently used device might become obsolete in
       a relatively short period.
   -   When working towards a database which involves data from different
       schemes, sources and regions (or on a larger scale even countries) it
       involves different devices and settings used for collecting data and for the
       identification process.

In this context it seems most appropriate to use a mixed and/or decentralized
model for data collection and identification.

Depending on the amount of data to be expected, a mixed model might involve
large computing power to be able to quickly give feedback to volunteers. The
need for a central location for either the primary identification process (e.g.
assigning a species to a recording), or the validation process, is no different for a
mixed or decentralized model.

Ideally – to facilitate different types of volunteers and/or different levels of
experience in performing species ID - it should be possible that volunteers can
send in only recordings and receive identification information in return (mixed
model), parallel to a facility to do and check their own identification
(decentralized model).
        It seems most appropriate however to organize both path ways (the
facility to send in recordings, as well as the facility to do and check one’s own
identification) centrally (per country, monitoring scheme or project).

When looking at the Netherlands, the NDFF is a central location where data
concerning bat (and other) species from various sources is made available to
different parties. Data is being added via ad-hoc observations (either
waarneming.nl, telmee.nl, lower governmental bodies [e.g. municipalities] or
consulting parties) and of structured survey and monitoring schemes (NEM).
     Ad-hoc data is to be validated, as using software may lead to
misidentifications in that data (of unknown magnitude). When dealing with data
that has been analysed for species identification with software, it is of paramount
importance that not only the outcome of the analysis, but also metadata of the
recording device and settings and of the software used are transferred. Without
that information a true validation is not possible and there will be no certainty
that the species identification has been done correct. Not being able to validate
data may in turn lead to data not being used and/or invalid information being
used.

                    15
From data to information

7 References

Barré, K., Viol, Isabelle Le, Julliard, R., Pauwels, J., Newson, S.E., Julien, J-P.,
     Claireau F., Kerbiriou, C. and Bas, Y. 2019. Accounting for automated
     identification errors in acoustic surveys. Methods in Ecology and Evolution,
     2019 00:1–18. https://doi: 10.1111/2041-210X.13198
Barlow K.E., Briggs P.A., Haysom K.A., Hutson A.M., Lechiara N.L., Pacey P.A.,
     Walsh A.L. and Langton S.D. (2015) Citizen science reveals trends in bat
     populations: the National Bat Monitoring Programme in Great Britain.
     Biological Conservation, 182, pp. 14-26.
Brigham, R.M., E.K.V. Kalko, G. Jones, S. Parsons & H.J.G.A Limpens, eds. 2004.
     Bat Echolocation Research: tools, techniques and analysis. Bat Conservation
     International, Austin, Texas. 167 pp.
Fenton, M.B, B.W. Keeley, J.D. Tyburec 2017. Learning to Listen. 2017 Bat
     Echolocation Symposium. 33 pp.
Loeb, S.C., T.J. Rodhouse, L.E. Ellison, C.L. Lausen, J.D. Reichard, K.M. Irvine,
     T.E. Ingersoll, J.T.H. Coleman, W.E. Thogmartin, J.R. Sauer, C.M. Francis,
     M.L. Bayless, T.R. Stanley, and D.H. Johnson. 2015. A plan for the North
     American Bat Monitoring Program (NABat). General Technical Report SRS-
     208. Asheville, NC: U.S. Department of Agriculture Forest Service, Southern
     Research Station. 112 p.
Mac Aodha O, Gibb R, Barlow KE, Browning E, Firman M, Freeman R, Harder, B.,
     Kinsey, L., Mead, G., Newson, S.E., Pandouski, I., Parsons, S., Russ, J.,
     Szodoray-Paradi, A., Szodoray-Paradi, F., Tilova, E., Girolami, M., Brostow
     ,G. and Jones, K.E. 2018. Bat detective—Deep learning tools for bat
     acoustic signal detection. PLoS Comput Biol 14(3): e1005995.
     https://doi.org/10.1371/journal.pcbi.1005995
Newman, C., Buesching, C.D. and Macdonald, D.W., 2003. Validating mammal
     monitoring methods and assessing the performance of volunteers in wildlife
     conservation- "Sed Quis Custodiet Ipsos Custodies?” Biological Conservation
     113: 189-197.
Newson, S.E., H.E. Evans & S. Gillings, 2015. A novel citizen science approach
     for large-scale standardised monitoring of bat activity and distribution,
     evaluated in eastern England. - Biological Conservation 191:38–49.
Roche, N., Catto, C., Langton, S., Aughney, T., Russ, J. 2005. Development of a
      Car-based monitoring protocol for the Republic of Ireland. Irish Wildlife
       Manuals, Vol. 19. National Parks and Wildlife Service, Department of
       Environment, Heritage and Local Government, Dublin, Ireland 42 Pp.
Roche, N. S. Langton , T. Aughney, J. M. Russ , F. Marnell , D. Lynn & C. Catto.
      2011. A car-based monitoring method reveals new information on bat
       populations and distributions in Ireland. Animal Conservation 14: 642–
       651.
Russo, D., and Voigt, C. C., 2016. The use of automated identification of bat
     echolocation calls in acoustic monitoring: A cautionary note for a sound
     analysis. Ecological Indicators, 66, 598–602.
     https://doi.org/10.1016/j.ecolind.2016.02.036

                   16
From data to information

7.1 Websites and web based –bat- data collection programs

Bat Detective: https://www.batdetective.org/ Citizen science for eco-acoustic
  biodiversity monitoring. Rory Gibb, Oisin Mac Aodha, and Kate E. Jones
  describe how improvements in technology are enabling the public to assist in
  monitoring global bat populations using sound. Environmental Scientist
  august 2016.
Bat Conservation International, n.d. Launch a global bat database. Retrieved
  2017-03-27, from http://www.batcon.org/our-work/initiatives/launch-a-
  global-bat-database
Bat Conservation Ireland, n.d. Car based bat monitoring. Retrieved 2017-03-27,
  from http://www.batconservationireland.org/what-we-do/monitoring-
  distribution-projects/car-based-bat-monitoring
Bat Conservation Trust, n.d. iBats. Retrieved 2017-03-27, from
  http://www.bats.org.uk/pages/ibatsprogram.html
Bat Conservation Trust, n.d. National Bat Monitoring Program. Retrieved 2017-
  03-27, from http://www.bats.org.uk/pages/nbmp.html
 GUANO [Grand Unified Acoustic Notation Ontology] is a universal, extensible,
  open metadata format for bat acoustic recordings.
  https://guano-md.org/
Norfolk Bat Survey, n.d. Norfolk Bat survey. Retrieved 2017-03-27, from
  http://www.batsurvey.org/norfolk/
Southern Scotland Bat Survey, n.d. Southern Scotland Bat survey. Retrieved
  2017-03-27, from http://www.batsurvey.org/scotland/
USGS, 2017 (January 5). The USGS Bat Population Data Project. Retrieved 2017-
  03-27, from https://my.usgs.gov/bpd/
USGS, 2016 (November, 2016). North American Bat Monitoring Program
  (NABat). Retrieved 2017-03-27, from https://www.fort.usgs.gov/science-
  tasks/2457 Wisconsin Bat Program, n.d. Acoustic Bat Monitoring. Retrieved
  2017-03-27, from http://wiatri.net/Inventory/Bats/volunteer/acoustic.cfm
Zoogdiervereniging & CBS: Dutch Mammal Society & Netherlands Central Bureau
  for Statsitics: acoustic car monitoring for selected bat species.
  http://www.zoogdiervereniging.nl/invoerpagina-zoogdier-meetnetten
  http://vleermuistransecttellingen.ndff.nl/

7.2 Other references
Adams, A.M. 2013. Assessing and analyzing bat activity with acoustic
     monitoring: Challenges and interpretations. University of Western Ontario –
     Electronic Thesis and Dissertation Repository, paper 1333.
Adams, M.A., M.K. Jantzen, R.M. Hamilton & M.B. Fenton 2012. Do you hear
     what I hear? Implications of detector selection for acoustic monitoring of
     bats. - Methods in Ecol. and Evol.3:992-998.
Arthur, L., M. Lemaire, L. Dufrêne, I. Le Viol, J.F. Julien & C. Kerbiriou, 2014.
     Understanding Bat-Habitat Associations and the Effects of Monitoring on
     Long-Term Roost Success using a Volunteer Dataset. Acta Chiropterologica
     Vol. 16 (2):397-411. doi: 10.3161/150811014X687350
Barlow K.E., Briggs P.A., Haysom K.A., Hutson A.M., Lechiara N.L., Pacey P.A.,
     Walsh A.L. and Langton S.D. (2015) Citizen science reveals trends in bat

                   17
From data to information

        populations: the National Bat Monitoring Programme in Great Britain.
        Biological Conservation, 182, pp. 14-26.
Barlow, K.E., Briggs, P.A., Haysom, K.A., Hutson, A.M., Lechiara, N.L., Racey,
      P.A., Walsh, A.L., Langton, S.D., 2015. Citizen science reveals trends in bat
      populations: the National Bat Monitoring Programme in Great Britain. Biol.
      Conserv. 182, 14–26.
Beeker, T.A., Millenbah, K.F., Gore, M.L., Lundrigan, B.L. 2013. Guidelines for
      creating a bat-specific citizen science acoustic monitoring program. Human
      Dimensions of Wildlife: An international Journal, Vol. 18, Is. 1, pp. 58-67
Biscardi, S. , J. Orprecio, M. B. Fenton, A. Tsoar, & J.M. Ratcliffe, 2004. Data,
      sample sizes and statistics affect the recognition of species of bats by their
      echolocation calls. - Acta Chiropterologica, 6(2): 347–363, PL ISSN 1508-
      1109 © Museum and Institute of Zoology PAS.
Brinkmann, R., O. Behr, I. Niermann & M. Reich (eds.), 2011. Entwicklung von
      Methoden zur Untersuchung und Reduktion des Kollisions-risikos von
      Fledermäusen an Onshore-Windenergieanlagen. - Umwelt und Raum Bd. 4,
      40-115, Cuvillier Verlag, Göttingen.
Burnett, S. C., and W. M. Masters. 1999. The use of neural networks to classify
      echolocation calls of bats. Journal of the Acoustical Society of America
      106:2189.
Clement, M.J., Murray, K.L., Solick, D.I., Gruver, J.C., 2014. The effect of call
      libraries and acoustic filters on the identification of bat echolocation. Ecol.
      Evol. 4:3482–3493
Corben, C., and G. M. Fellers. 2001. Choosing the “correct” bat detector—a reply.
      Acta Chiropterologica 3(2):253–256.
Duffy, A. M., L. F. Lumsden, C.R. Caddle, R.R. Chick, and G.R. Newell. 2000. The
      efficacy of Anabat ultrasonic detectors and harp traps for surveying
      microchiropterans in south-eastern Australia. Acta Chiropterologica, 2: 127–
      144.
Fischer, J., J. Stott, B. S. Law, M. D. Adams, and R. I. Forrester. 2009. Designing
      effective habitat studies: quantifying multiple sources of variability in bat
      activity. Acta Chiropterologica, 11: 127–137
Frick, W.F., 2013. Acoustic monitoring of bats, considerations of options for long-
      term monitoring. - THERYA, Vol.4(1):69-78. DOI: 10.12933/therya-13-109
Froidevaux, J. S. P., Zellweger, F., Bollmann, K., & Obrist, M. K., 2014.
      Optimizing passive acoustic sampling of bats in forests. Ecology and
      Evolution, 4(24):4690–4700. http://doi.org/10.1002/ece3.1296
Hawkins C., Browning E. and Jones K.E. (2016) iBats Jersey - Review: Report for
      States of Jersey Department of the Environment. Centre for Biodiversity and
      Environment Research, University College London.
Herr, A., Klomp, N.I., and Atkinson, J.S. 1997. Identification of bat echolocation
      calls using a decision tree classification system [online]. Complex. Int. 4.
      Available from
      http://journalci.csse.monash.edu.au/ci_louise/vol04/herr/batcall.html
      [accessed 1 April 2008].
Jansen, E.A., H.G.J.A. Limpens, J.J.A. Dekker, M. Liefting & T. van der Meij,
      2012. Monitoring of bat species currently not covered by the Dutch national
      monitoring scheme. Volunteers, design & statistical power. Report 2012.04.
      Zoogdiervereniging, Nijmegen.

                   18
From data to information

Jones G., Jacobs D.S., Kunz T.H., Willig M.R. and Racey P.A. (2009) Carpe
     noctem: the importance of bats as bioindicators. Endangered. Species
     Research, 8, pp. 93-115.
Jones K.E., Russ J.A., Bashta A., Bilhari Z., Catto C., Csosz I., Gorbachev A.,
     Gyorfi P., Hughes A., Ivashkiv I., Koryagina N., Kurali A., Langton S., Collen
     A., Margeian G., Pandourski I., Parsons S., Prokofev I., Szodaray-Paradi A.,
     Szodaray-Paradi F., Tilova E., Walters C.L., Weatherill A. and Zavarzin O.
     (2013) The Indicator Bats program: A system for the global acoustic
     monitoring of bats. In: Collen B., Pettorelli N., Baillie J., Durant S. (eds.)
     Biodiversity Monitoring and Conservation: Bridging the Gap between Global
     Commitment and Local Action. Wiley-Blackwell, United Kingdom. 6.
     iBats, http://www.ibats.org.uk.
Kerbiriou, C., J.F. Julien, K, Ancrenaz, A.S. Gadot, G. Loïs, F. Jiguet & R. Julliard,
     2008. Suivi des espèces communes après les oiseaux ... les chauves-souris?
     Muséum National d'Histoire Naturelle, Unité Conservation des Espèces
     Restauration et Suivi des Populations. 6pp.
Lefevre, A., Verkem, S., Onkelinx, T. & Verbeylen, G. 2008. Zoogdieren in de
     stad: vleermuizenmonitoring met behulp van voertuigtransecten. Rapport
     Natuur.studie 2008/13, Natuurpunt Studie (Vleermuizenwerkgroep),
     Mechelen, België.
Lemen, C., Freeman, P.W., White, J.A., Andersen, B.R., 2015. The problem of
     low agreement among automated identification programs for acoustical
     surveys of bats. West. N. Am. Nat. 75, 218–225
Limpens, H.J.G.A., K. Mostert & W. Bongers, 1997. Atlas van de Nederlandse
     vleermuizen; onderzoek naar verspreiding en ecologie. - KNNV Uitgeverij,
     260 pp.
Limpens, H.J.G.A. & A. Roschen, 1996. Bausteine einer systematischen
     Fledermauserfassung, Teil 1: Grundlagen. - Nyctalus (N.F.) 6, Heft 1, S. 52-
     60.
Limpens, H.J.G.A. & A. Roschen, 2002. Bausteine einer systematischen
     Fledermauserfassung. Teil 2 - Effektivität, Selektivität, und Effizienz von
     Erfassungsmethoden. Nyctalus (N.F.) 8/2:155-178.
Limpens, H.J.G.A., M. Boonman, F. Korner-Nievergelt, E.A. Jansen, M. van der
     Valk, M.J.J. La Haye, S. Dirksen & S.J. Vreugdenhil, 2013. Wind turbines
     and bats in the Netherlands - Measuring and predicting, toepasbaarheid
     voor de nederlandse windsector. Zoogdiervereniging & Bureau
     Waardenburg. 3pp.
Limpens, H.J.G.A. E.A. Jansen, L. Höcker & M. Schillemans, 2016. Monitoring of
     Bats in an Urban Landscape - A monitoring system for bats in urban
     landscapes in the framework of the assessment of their conservation status
     (FCS). Rapport 2015.023. Bureau van de Zoogdiervereniging, Nijmegen.
Loeb, S.C., T.J. Rodhouse, L.E. Ellison, C.L. Lausen, J.D. Reichard, K.M. Irvine,
     T.E. Ingersoll, J.T.H. Coleman, W.E. Thogmartin, J.R. Sauer, C.M. Francis,
     M.L. Bayless, T.R. Stanley, and D.H. Johnson. 2015. A plan for the North
     American Bat Monitoring Program (NABat). General Technical Report SRS-
     208. Asheville, NC: U.S. Department of Agriculture Forest Service, Southern
     Research Station. 112 p.
Lundy, M., E. Teeling, E. Boston, D. Scott, D. Buckley, P. Prodohl, F. Marnell and
     I. Montgomery. 2011. The shape of sound: elliptic Fourier descriptors (EFD)

                    19
From data to information

     discriminate the echolocation calls of Myotis bats (M. daubentonii, M.
     nattereri and M. mystacinus). - Bioacoustics Vol. 20(2).
McLellan R., Iyengar L., Jeffries B. and Oerlemans N. (2014) WWF Living Planet
     Report. WWF, Gland, Switzerland.
Meyer, C.F.J., 2015. Methodological challenges in monitoring bat population- and
     assemblage-level changes for anthropogenic impact assessment. Mamm.
     Biol. 80, 159–169.
Milne, D. J., M. Aarmstrong, A. Fisher, T. lores, and C. R. Pavey. 2004. A
     comparison of three survey methods for collecting bat echolocation calls and
     species-accumulation rates from nightly Anabat recordings. Wildlife
     Research, 31:57–63.
Newson, S.E., Ross-Smith, V., Evans, I., Harold, R., Miller, R., Barlow, K., 2014.
     Bat-monitoring: a novel approach. Br. Wildl. 25, 264–269.
O’Shea, T.J. and Bogan, M.A., eds., 2003, Monitoring trends in bat populations of
     the United States and territories: problems and prospects: U.S. Geological
     Survey, Biological Resources Discipline, Information and Technology Report,
     USGS/BRD/ITR--2003–0003, 274 p
Sattler, T. 2009. Biodiversity in urban landscape matrices: from species richness
     to functional community structure. PhD Thesis, University Bern.
Verboom, B. & H.J.G.A. Limpens. 2004. Methodieken verspreidingsonderzoek
     landzoogdieren van de inhaalslag. Rapport VZZ 2004.12 in opdracht EC-
     LNV. 64 pp.

                   20
From data to information

8 Bijlages

Appendix I: Questionnaire

                  21
From data to information

I) Appendix I: Questionnaire

Instruction for the questionnaire of the NLBIF project dealing with automatic
generated data on bat acoustics.

The project for which we are looking for your input and cooperation is dealing
with how to handle the data stream generated through the projects working with
auto acoustic recording and auto ID of acoustic bat signals.

Please answer the questions we posed in the different tabs in this Excel table
questionnaire. Many questions are written in the so called ‘closed’ format. We
would like to invite you however, to answer in full text. Also we would like to
invite you to point out topics we are missing.

The information needed to try and improve the handling of the data stream of
such bat research projects, is primarily connected to the topics:

  A metadata (Table 4)
  B data retrieval, management and storage (Table 5)

So please address the questions in their specific excel-tabs.

In case you are not the right person to fill in the excel tabs, would you please let
us know whom we should address with this first questionnaire and possible more
detailed questionnaires?

Please let us know whether we could contact you again to look at some more
detailed questions in a next phase of the project.

If you know of any published literature dealing with data handling bat data could
you pls. supply the reference
We supply extra tabs dealing with details. It might be helpful to also try and
answer the questions in the excel-tabs dealing with:

  C recording bats
  D species ID, and
  E observation - monitoring design

These tabs are not required the primary aim of our project, but provide useful
further insights. We leave it at your discretion whether or not you fill out these
tabs.

                   22
From data to information

Table 4. Tab A of questionnaire: metadata

                                                                     Please indicate
WHICH METADATA ARE RECORDED, OR DEDUCIBLE FROM
                                                                     correct choice
RAW DATA?
                                                                     by background
                                                                     colour
Metadata
Are you recording and storing metadata?                              Yes               No
Are you using a custom made format, for encoding of metadata
?                                                                    Yes               No
Are you using a freely available format?                             Yes               No
Other,….

Date and time of recording?                                          Yes               No

Location coordinates?                                                Yes               No

Metadata recording                                                   Yes               No
  Is this information available from your software?                  Yes               No

Registered weather data                                              Yes               No
  Is this information available from your recording?                 Yes               No

Data on equipment & observer                                         Yes               No
Are you recording and storing ..
  specifics and ID of your recording unit ?                          Yes               No
  name observer (+ observer team)?                                   Yes               No
  specific of material (incl. mic and sensitivity, and settings of
automatic detectors)                                                 Yes               No

Other,….
Metadata on sample point                                             Yes               No

Metadata on sample session / recording session                       Yes               No

Metadata on ID-proces                                                Yes               No
Are you recording and storing ..
specifics of identifying person/party/software                       Yes               No
Are you recording and storing ..                                     Yes               No
  specifics on call quality?                                         Yes               No
  specifics on type/version/year of classifier?                      Yes               No
  specifics on used filter settings?                                 Yes               No
  used reference material?                                           Yes               No
Other,….

How do you link metadata - (ultra)sound recording - raw data and
processed data?

                      23
From data to information

24
From data to information

Table 5. Tab B of the questionnaire: data retrieval and storage

                               Please
DATA COLLECTION,               indicate
RETRIEVING DATA FROM           correct
PARTICIPANTS AND DATA          choice by
STORAGE                        background
                               colour           correct choice!

Does the data concern a
specific project or multiple
projects?                      Specific         Multiple
Does the data concern a
single observer/party or
multiple observers/parties     Single           Multiple
Is there a facility to
reanalyse data?                Yes              No
Other, ……..

With wat technique,
process and in which form
are data retrieved ..
                                                                    File transfer
                                                (physical           (e.g.
How is the data transferred?
                                                exchange of)        wetransfer,
                               Portal           Storage Device      shared disks)
Other, ……..

                                                Scrubbed Raw
                                                (all noise and
hat data is transferred?       Raw Only         non-target
                               (e.g. all        species data
                               recordings)      removed)            ID only           ID and recordings
Are the potential biological
sounds in the cast away
retrieved and stored?          Yes              No
Other, …

If scrubbed data is used
                               Before
                               retrieval,
                               by
At what stage and by what
                               observer,
party is the data scrubbed?
                               facility (e.g.   Before retrieval,   After retrieval
                               software)        by observer, no     of data
                               provided         facility provided   (centrally)
Other, ……..

Validation
Validation done?                Yes             No
       Is all data validated or
a sample of it?                 All             Sample
Done by observers?             Yes              No

                      25
From data to information

         If done by observers:
is facility (e.g. software)
provided by you?                  Yes        No
         If facility is provided:
does the facility operate
locally (e.g. downloaded
software) or on network (e.g.
internet)                         Locally    Network

Done after retrieval
(centrally)?                     Yes         No
       If done centrally: only
via software, only by hand
vetting or by software and       software
hand vetting                     only        manually only   combination

Storage
Are you storing data
(including and especially raw
data e.g. recordings) on
(local or SAN) hard disks or
in the cloud?                    SAN/Local   Cloud
What was the reason for this
choice?
Remarks

Who is presently managing
the collated data?
How is future management
of the collated data secured?
Other, ….

Who is providing storage
capacity for you/for your
project
How is future storage
capacity for your data
secured?
Other, ….

What is the average storage
capacity you currently need
for e.g. 1 year of data?
Can you give an estimate of
how this required capacity
will develop in the near
future?
Other, ….

What technique do you use
to retrieve stored data?
How many times/year do you
retrieve stored data?

                       26
From data to information

II)   Appendix II

                27
From data to information
Summary interpretation of questionnaire acoustic auto-data bats
8 respondents – not always everybody responded to every question. Dutch Mammal Society excluded in
below summary

 WHICH META DATA ARE RECORDED, OR DEDUCIBLE FROM RAW DATA?

 Meta data                                              yes
 Are you recording and storing meta data?                7/8   7/8 store meta data
 Are you using a custom made format, for encoding of           4/8 use a custom made; those who don't use custom made use
 meta data?                                             4/8    freely available format
                                                               5/8 use freely available format; 1/8 uses both custom and freely
 Are you using a freely available format?
                                                        5/8    available

 Date and time of recording?                            1/8    1/8 doesn't store date and time; why?

 Location coordinates?                                  8/8    8/8 store location coordinates

                                                               4/8 store meta data of the recording itself; not everyone
 Meta data recording                                           understands what metadata on recording is/our question needs
                                                        4/8    definition
   Is this information available from your software?    4/8    4/8 have data available from software

 Registered weather data                                6/8    6/8 store weather data; question needs definition
                                                               3/8 store from recording; 1/8 collects his own data; 1/8 uses data
   Is this information available from your recording?   6/8    from weather station

 Data on equipment & observer                           5/8    5/8 store data on equipment
 Are you recording and storing..
   Specifics and ID of your recording unit?             8/8

                   2
You can also read